Shell step-by-step monitoring of iostat data and shell step-by-step iostat

Shell step-by-step monitoring of iostat data and shell step-by-step iostat

In performance testing, iostat monitoring data is too bad to display and looks very painful, such

Write a shell script to make the changes as follows:

#! /bin/bashinterval=1if [ $# -gt 0 ]; then  interval=$1fiwhile (true) do    iostat -k; sleep $interval; done | awk 'BEGIN{    data_flag = 0;        sum_tps = 0;    sum_read_kb_speed = 0;    sum_write_kb_speed = 0;    sum_read_kb = 0;    sum_write_kb = 0;        }{        if ( $1 == "Device:")    {        data_flag = 1;        #print "+ " $0            }    else if ( NF == 0 )    {        if (data_flag == 1)        {            printf("%s tps %6.2f kB_read/s %10.2f kB_write/s %10.2f kB_read %14d kB_write %14d\n",                 strftime("%D %H:%M:%S"),                sum_tps,                sum_read_kb_speed,                sum_write_kb_speed,                sum_read_kb,                sum_write_kb);                data_flag = 0;                sum_tps = 0;            sum_read_kb_speed = 0;            sum_write_kb_speed = 0;            sum_read_kb = 0;            sum_write_kb = 0;                        #print "- " $0        }        else        {            #print "= " $0        }        #data_flag=0;    }    else    {        if (data_flag == 1)        {            #print "1 " $0;                        tps = $2;            read_kb_speed = $3;            write_kb_speed = $4;            read_kb = $5;            write_kb = $6;                        sum_tps += tps;            sum_read_kb_speed += read_kb_speed;            sum_write_kb_speed += write_kb_speed;            sum_read_kb += read_kb;            sum_write_kb += write_kb;                                }        else        {            #print "0 " $0;        }    }    }END{}    

The effect after running is as follows:

Is the display much better. There is time to display, so it is much easier to locate the problem after logging.

 

For iostat reference, see: http://www.ha97.com/4546.html